Hello all,

I am still having a serious issue with respect to the time it takes
retrospect to go around my network and pick up all the clients. I have
about 60 machines to back up every night and I configured my script to
backup only the changes from the previous night. I have "Save Source for
Snapshots", , "Match Source Volumes to Catalog" and "Don't add
duplicates to Backup Set" enabled in my script's options. I also keep
the catalog compressed. My current backup set has 7 members (tapes) and
the catalog file (compressed) is about 142MB.
The problem is that Retrospect spends over 90% of the time "Matching"
files from the clients to the catalog and very little time backing up
(writing to the tape). My network throughput is not fantastic, yet not
bad. I noticed that during the "Matching" period my CPU usage is 100%.

I would like to shorten as much as I can this "matching" time and use my
14 hours allotted to backing up for doing the actual work. Does anybody
know what can I do software/hardwarewise to achieve this?

Here is my hardware: PII266, 196MB RAM, 6.4GB HDD, Quantum DLT 4000 SCSI
Tape Drive.

Here is a copy of my latest backup log stats:

 "8/28/2000 2:02:04 AM: Copying LOCAL DATA (Y:) on SOLITON
  8/28/2000 2:02:04 AM: No files need to be copied
  8/28/2000 2:14:44 AM: Snapshot stored, 3,034 KB
  8/28/2000 2:15:18 AM: Execution completed successfully
  Duration: 00:12:50 (00:09:47 idle/loading/preparing)

  8/28/2000 2:28:15 AM: Execution incomplete
  Total performance: 7.8 MB/minute
  Total duration: 15:42:04 (15:01:45 idle/loading/preparing)
  Quit at 8/28/2000 2:29 AM"

Dantz, feel free to jump in and offer solutions to this issue if there
is a known issue or if you guys have a solution.

Thanks, Ben

Ben Mihailescu
System Administrator
Electrical and Computer Engineering, McGill University

To subscribe:    [EMAIL PROTECTED]
To unsubscribe:  [EMAIL PROTECTED]
Archives:        <http://list.working-dogs.com/lists/retro-talk/>
Problems?:       [EMAIL PROTECTED]

Reply via email to